A Förster Resonance Energy Transfer Switchable Fluorescent Probe With H2S-Activated Second Near-Infrared Emission for Bioimaging

Real-time and accurate detection of endogenous hydrogen sulfide is of great biomedical significance. Here, a FRET-based fluorescent probe for ratiometric detection of H2S was designed to comprise an AIE luminophore TPE as an energy donor and a monochlorinated BODIPY dye as an energy acceptor and H2S-responsive site. Such a designed probe showed H2S-dependent ratiometric and light-up NIR-II emission, enabling accurate imaging of H2S-rich cancer cells and identification of H2S-rich tumors with high resolution.


INTRODUCTION
Endogenous hydrogen sulfide is an important signaling molecule, mainly derived from the enzymatic hydrolysis of L-cysteine Singh et al., 2009). Studies have found that H 2 S is associated with many pathological processes, while an abnormal level of H 2 S may associated with some diseases, such as Alzheimer's disease, hypertension, and cardiac ischemia disease (Eto et al., 2002;Zhao et al., 2015;Shi et al., 2017). Therefore, real-time and accurate detection of hydrogen sulfide is of great biomedical significance. Until now, many fluorescence-based H 2 S probes have been reported (Jin et al., 2016;Wang et al., 2018Wang et al., , 2019b; however, the fluorescence of many probes generally locates in the visible or the near-infrared I region (650-900 nm), inevitably leading to some drawbacks of poor tissue penetration, severe background interference from living tissue (Zhou et al., 2014). Compared with the traditional NIR-I imaging (650-900 nm) (Li et al., 2018), fluorescent imaging in the second near-infrared window (NIR-II, 1,000-1,700 nm) has attracted more and more attention due to lower tissue absorption, stronger tissue penetration, and reduced autofluorescence (Hong et al., 2012;Dang et al., 2016;Shi et al., 2018;Xu et al., 2018). Another issue is the hydrophobic nature of most traditional fluorescent dyes, which generally triggers the aggregation in physiological conditions due to π-π stacking. Such a process can give rise to aggregation-caused quenching (ACQ) (Sun et al., 2014;Yuan et al., 2015) and thus compromise the accuracy of bioimaging. In comparison, fluorogens with AIE characteristics show enhanced fluorescence in the aggregate states, thus providing an alternative strategy for the design of fluorescent light up probes (Zhao et al., 2012;Mei et al., 2014Mei et al., , 2015Zhang et al., 2015;Fu et al., 2019). Since the hydrophobic fluorescent probes undergo the intrinsic aggregation process in aqueous media, it is desirable to develop H 2 S-activatable probes with AIE characteristic for in vivo imaging.
Herein, we reported a H 2 S-responsive probe that showed ratiometric fluorescence and NIR-II emission light-up upon activation for in vitro and in vivo imaging (Scheme 1). Such a probe was designed by appending an AIE luminophore TPE, as an energy donor, to a monochlorinated BODIPY dye as an energy acceptor and H 2 S-responsive site. As compared to conventional intensity-based fluorescent probes (Huang et al., 2014;Tang et al., 2016), this Förster resonance energy transfer (FRET)-based ratiometric probe can enable accurate detection through elimination of the limitations of experimental conditions including probe concentration, light source, and background SCHEME 1 | Schematic illustration of the design of probe TPE-BODIPY-Cl and the mechanism for H 2 S-mediated ratiometric and NIR fluorescence light up. SCHEME 2 | Synthesis of the target compound TPE-BODIPY-Cl.
interference effects (Wang et al., 2019a). Such a design strategy is applicable to the design of various ratiometric probes for different targets. As expected, in the absence of H 2 S, due to the good spectral overlap between the emission spectra of the TPE and the absorption spectra of the BODIPY, efficient FRET occurs. In contrast, in the presence of H 2 S, the absorption spectra of BODIPY undergo an obvious red shift, resulting in a significant reduction of the overlap with the TPE emission. Correspondingly, the FRET process is significantly attenuated. More importantly, upon activation by H 2 S, the probe produces a new fluorescence light-up at 920 nm with the fluorescence tail extending to 1,300 nm, indicative of the suitability for fluorescent imaging in the second near-infrared (NIR-II).

EXPERIMENTAL Synthesis
The TPE-BODIPY-Cl was obtained from the synthetic route of Scheme 2. Br-TPE and BODIPY were synthesized according to the literature procedure (Zhao et al., 2013(Zhao et al., , 2014. Animal experiments were performed in compliance with Chinese legislation on the Use and Care of Research Animals and guidelines by Fudan University Animal Studies Committee for the Care and Use of Laboratory Animals. All experimental procedures were approved by this committee.

Synthesis of Compound A
Br-TPE (165 mg, 0.39 mmol) and 2,3,3-trimethyl-3H-indole (62 mg, 0.39 mmol) were dissolved in 25 mL CH 3 CN and refluxed for 10 h. Then, the solvent was removed under reduced pressure, and the crude product was dissolved in CH 2 Cl 2 . The mixture was dropped into the ether solvent to precipitate white solid, which was used for next reaction without further purification. HRMS (ESI, m/z): calculated for C 38 H 34 N [M-Br] + : 504.2691, found: 504.2699.

Spectroscopic Studies of TPE-BODIPY-Cl
With the probe in hand, we initially evaluated the photophysical properties. Because the probe contains the TPE AIEgen, we explored the AIE performance of the probe to obtain the best test conditions. As shown in Figure 1A, we tested the FRET process of the probe under different ratios of H 2 O/CH 3 CN. With the increasing water content, the degree of aggregation of the probe intensifies, accompanying the increase of TPE fluorescence while the occurrence of ACQ for BODIPY chromophore. As is well-known, a ratiometric fluorescence mode has higher accuracy than turn-on or turn-off fluorescence detection mode Zhang et al., 2019), we here selected Tris/CH 3 CN buffer solution (0.5 M Tris, 40% CH 3 CN, pH = 7.4) as the next testing condition in order to obtain the ratiometric fluorescent responsiveness. The aggregation of our probe under this buffer solution was proven by dynamic light scattering ( Figure S1). Next, we evaluated the response capability of TPE-BODIPY-Cl toward H 2 S (Figure 1 and Figure S2). As shown in Figure 1B, the free probe showed strong absorption at 550 nm. The typical absorption band of the TPE around 300-360 nm was also noted. In the fluorescence spectrum, due to FRET process, we can observe two strong fluorescence peaks with maxima at 438 and 598 nm, corresponding to TPE and BODIPY, respectively. When treated with 100 µM H 2 S, the absorption band at 550 nm decreased significantly and a new absorption band appeared at 760 nm with a red-shift of 220 nm due to the formation of TPE-BODIPY-SH that was proven by HRMS analysis (Figure S3). Such treatment with NaHS attenuated the FRET, thus affording an enhancement of the fluorescence intensity ratio (I 438 /I 598 ) by 12 times. This indicated that TPE-BODIPY-Cl was indeed a ratiometric fluorescent probe for H 2 S. Most importantly, H 2 S-triggered a new NIR-II fluorescence light up with a maximum emission of 920 nm (λ ex = 760 nm). These results demonstrated that TPE-BODIPY-Cl could be used as a H 2 Sactivatable NIR-II fluorescent probe to enable high-resolution bioimaging with deep-tissue penetration. Utilizing the linear relation of fluorescence intensity ratio at 438 and 598 nm with H 2 S concentration (0-50 µM) (Figure S4), the detection limit was determined to be 6.5 × 10 −7 M, indicating that TPE-BODIPY-Cl has high sensitivity for H 2 S detection. Of note, the probe exhibits minimal optical responsiveness to biologically related reactive sulfur (RSS), oxygen (ROS), and nitrogen species (RNS) and some ions, showing its high selectivity for H 2 S ( Figure 1F and Figure S5). In addition, the good photostability of probe TPE-BODIPY-Cl, evidenced by minimal optical changes under continuous irradiation with light irradiation (Figure S6), indicated its suitability for bioimaging.

Imaging of H 2 S in Living Cells
Inspired by the promising response to H 2 S, we then assessed the ability of TPE-BODIPY-Cl for fluorescence imaging in living HCT116 cells that express high levels of H 2 S [20-100 µM, Szabo et al., 2013]. As shown in Figure 2, the incubation of HCT116 cells and 10 µM TPE-BODIPY-Cl for 30 min afforded the bright and stable fluorescence signal in the green channel and relatively weak fluorescence in the red channel. The ratio of the green to red channel is ∼2.82. When a CBS inhibitor aminooxyacetic acid (AOAA) which can inhibit the H 2 S production was added,  the ratio dropped to 0.70. In contrast, with the addition of an allosteric CBS activator S-adenosyl-L-methionine (SAM) to promote the production of H 2 S, the ratio increased to 3.59. These results indicated that TPE-BODIPY-Cl can efficiently enter living cells and serve as a potential sensor to detect endogenous hydrogen sulfide rapidly and specifically.

Imaging of H 2 S in vivo
Finally, we explored the ability of the probe for visualizing H 2 S-rich cancers using HCT116 subcutaneous xenograft nude mice. TPE-BODIPY-Cl was administrated to nude mice through intratumoral injection. As shown in Figure 3, after the injection, obvious NIR-II fluorescence in the tumor region was observed and the signals gradually increased over time, producing a 14.8fold enhancement at the time point of 60 min ( Figure S7). These results indicated that the TPE-BODIPY-Cl could be activation of NIR-II fluorescence in H 2 S-rich colorectal cancers.

CONCLUSION
In summary, we have designed a FRET based probe through appending the AIE luminophore TPE to the monochlorinated BODIPY dye for imaging of H 2 S-rich cancer cells and tumors, wherein TPE serves as an energy donor and BODIPY dye as an energy acceptor. This probe showed H 2 S-dependent FRET process, thus enabling the selective visualization of endogenous H 2 S in HCT116 cells. Furthermore, this probe displayed H 2 S specific activation of NIR II emission light up. By using this activatable NIR II emission, accurate identification of colorectal tumors was realized. We expect our design strategy here can help the development of a new activatable probe.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this manuscript will be made available by the authors, without undue reservation, to any qualified researcher.

ETHICS STATEMENT
The animal study was reviewed and approved by Fudan University.

AUTHOR CONTRIBUTIONS
CZ and XG conceived the project and wrote the manuscript. RW conceived the molecule design. RW and WG prepared and characterized the small molecule. RW, TZ, and GX performed the optical characterization. RW and JG performed the living cells imaging, in vivo imaging, and analyzed the data.